A new analysis has shown that TikTok's algorithmic systems enabled the dissemination of ethnic hate speech and fabricated political content during Kenya's 2022 election cycle, with videos reaching over 4 million users.
Historical Echoes in Digital Disinformation ¶
Research by Mozilla fellow Odanga Madung, titled How Disinformation on TikTok Gaslights Political Tension in Kenya, examined 130 videos from 33 accounts that exploited TikTok's For You page algorithm to spread inflammatory content. The study found these videos used graphic imagery from the 2007/2008 post-election violence to stoke fears among Rift Valley communities.
"Many videos contained explicit threats of ethnic violence targeting Rift Valley communities, echoing the 2007/2008 conflict that killed over 1,000 people," the report states. The content included derogatory terms like 'madoadoa'—a label banned by the National Cohesion and Integration Commission in April 2022—to dehumanize specific groups.
Fabricated Media and Moderation Challenges ¶
The disinformation campaign also involved doctored news headlines from major Kenyan publications like the Daily Nation and The Standard, as well as fake KTN news bulletins. These materials falsely alleged scandals involving political figures, according to the report.
TikTok's policies prohibit hate speech and synthetic content, yet the platform failed to remove these violations. A former content moderator cited context bias, explaining that language barriers and rapid review processes hindered effective moderation. "I had to moderate Hebrew videos without understanding the language," the moderator told Mozilla.
Legal Frameworks and Platform Accountability ¶
Kenya's 2018 Computer Misuse and Cybercrimes Act criminalizes hate speech, but legal experts note enforcement challenges. Abdulmalik Sugow of the Kofi Annan Foundation warned that global platforms like TikTok operate under foreign regulatory frameworks, raising concerns about transparency and accountability.
Previous research conducted by Mozilla in 2021 exposed how Twitter was weaponized by foreign actors to manipulate Kenya’s public discourses, echoing tactics linked to the 2017 Cambridge Analytica scandal. "Hateful rhetoric which may qualify as hate speech under the Kenyan Law could remain under the purview of social media platforms which have continued to regulate such content rather opaquely," Sugow said.
Platform Response and Ongoing Concerns ¶
TikTok removed some flagged content after the study's release, but Madung emphasized the importance of understanding how social media shapes political discourse. "This report highlights the critical role of global platforms in Kenya's information ecosystem," he said.
